Impact of Data-Driven on Digital Transformation

Fernando Ultremare

CTO at CINQ

Learning Objectives

A data-driven approach allows transforming data into powerful insights to guide the development and growth of successful digital products. This presentation will show a practical approach of how your company can create a growth strategy to evolve digital products, building a strong connection between product and business. Discover methods and practices to foster, strengthen and scale the data-driven culture in your company.


Key Takeaways:



  • Data-driven culture and digital transformation

  • Turn data into a valuable asset

  • Drive product results through continuous learning practices


"We start to run into think about our sprints in our product backlog as a series of experimentations instead of a series of predefined features."

Fernando Ultremare

CTO at CINQ

Transcript

Hi there, my name is Fernando Ultremare, and I am the CTO of CINQ. I’m going to talk about the impact of data driven on digital transformation, more specifically, how we can do better digital product development. I’m really happy to be here at CIO VISIONS to talk to you. I hope you enjoy this session. If you want to connect with me after the session, feel free to hit me using this handle.


Let’s start by telling you a little story that happened to me. It was the beginning of my career as an Agile Product and Software Project Manager. It was also the beginning of the Agile adoption at large here in Brazil. Everything was really immature at the time. We had this customer, they were a really important customer for us, one of the biggest accounts. We have been working for them for like eight months. After eight months, we had nothing [unintelligible] deliver it to production. The customer started to complain about it, ahe boss of our stakeholders started to question him about the value that the team was generating. We tried to create ideas to show his boss what we’re working on.


Out of the blue, they came up with this idea that we were supposed to print all the source code that we were developing. I tried to explain to them that it was not the case to have all the source code printed with no success. Since they were reading part of the customer, I ended up with two bags full of printed handouts like the ones you see here in this picture, in a flight, going to show our customer. I was really sorry by the trees that needed to be cut in order for us to show your customer that we’re actually doing something. But you wouldn’t be surprised by the number of teams and the number of companies that still do lots of metrics. They measure the process only using metrics that are only related to the size of the product. It doesn’t have any relation to the business impact there are creating.


For instance, things like the planning pokers, fibonacci stuff, or the tea size estimates, or even functional points are broadly used by companies in teams. Here at QINQ, we started to think that we needed to do this better. Many years ago, we started on this question of creating better ways to help our customers to deal with this product lifecycle. We strongly believe that if we can have an approach that is strongly based on data, we will create a sense of direction for our products in alignment, which is very important for all the teams that are working for the products. With that direction in alignment, we have the speed because we’re going with confidence, and also the ability to work at scale. So this is the most important impact of data driven on digital transformation because it doesn’t matter if you go fast if you’re going to the wrong direction.


There’s lots of things to think about here in terms of digital transformation, because it’s a really big and really broad concept. I’ve chosen the digital product. Part of it because they play a really important role in the digital transformation of our companies. The digital products are used by our customers to get the value that we generate from our business, or most of the business nowadays do lots of transactions through the digital products. Finally, all the team members, the employees of our company are related in the process of developing the digital products. So if we can have a better way, a modern In a way, a more digital, and native way of doing digital products, for sure we’re going to impact in a really positive way the overall digital strategy of the company, making it more digital.


To create this growth driven strategy, you need to combine these three pillars. The first one, as I already said, is the business indicators. Then, you need to understand the customers and combine the understand of the customers, their journeys, their sentiments, their experience with all the data that your systems are currently generating, and our coding currently available for a company. Then, we can follow this simple process, where we start to create a much more data driven approach to our DevOps development lifecycle that we went, ended up with a new kinds of backlog items to our product lifecycle, and I started from the data.


To start that—this is not that new—we started shifting our mindset from a project mindset to a product mindset. We started to create something that we call growth driven development. To start here is really important because, in a project, you are always trying to find or to create this long features roadmap planning, months and even year ahead the roadmap of our products. We have lots of top down backlog items that come from different stakeholders in the company, high level people, and this strong and really focus on tight schedules, and on the deliverables, the features, the actual stories that you’re developing. When you go to the product mindset, we start to think about the indicators of the business, the KPIs, the performance indicators. You try to validate your backlog before you start to develop in them or to put more effort on them. And your hours, always think thinking about hypothesis, and not certainties. So you always have this immediately to think about hypothesis, to check your hypothesis, to see if you’re going to the right direction in terms of the business results that you want to achieve.


07:38

So first, you need to start from the data. As I will show you later in this presentation, there are plenty of data already available in our companies. Like transactions, analytics, life’s online reviews, and our data science team work together or put this data together and inspire us to add techniques to analyze and to see this data from different angles. We put a layer of a user experience expert over that, and they collaborate together—the data science and the user experience expert—to create and learn about the user behavior. Then, in conjunction with the product manager, the business owners, and other people in the team, they start to think about problems and ideas to trigger experiments that we want to validate with new data. This is really important to work with this kind of experiments in your projects, because with that, we can actually and definitely create this culture of learning inside our companies, and learning from facts that we analyze, we generate from the experimentations that we run, okay. From that fact and from that understanding, we can have insights, we can have ideas to generate more triggers to the next experiments or the next ideas of features or functionalities that we need to add or we want to add to our products.


09:24

To give you here, a quick example of this in practice on a real use case, and I will deep dive into this customer in the end of the presentation.We have this customer, it’s a big customer. They have millions of mobile users on their mobile app. The company really wanted to increase the conversion rates of the users that downloads their applications or go to their website. The data science and the user experience experts, they were working together, analyzing the data from the reviews of the Apple Store and the Play Store, and combining this data with the analytics information of the user using the app. They could understand that it was probably the case that we have extra steps in the onboard process that were redundant, or they didn’t need to be put there at the time. They create this idea of an experiment of changing a little bit the onboarding process. They talked to the product manager, and he agreed to put a little story to create an AB test in the next sprint. With the data that they created from this experiment, they were able to confirm that they could increase in 3% the conversion rate of the system by doing just a small change in the application.


11:08

So what we want to be doing is to be running this cycle of experiments for the whole lifecycle of the products that we are developing. Always looking for the next indicator to be impacting, and sharing this view of the business with everyone with lots of people so they can engage into and be aligned with the with the proposition of the product. After that, we can collect the results, and also create amazing and awesome experience for our customers. So this was the first part of the session.


11:55

Now, I’m going to show you a practical approach how you can start just as simple. You can simply start that in a running product that you are already have in our company. You don’t need to do a big upfront project or infrastructure to start doing this kind of approach to an existing product that you are developing right now. At first, we try to help our customers to acknowledge that they already have valuable data available on their systems. For instance, on platforms they use, like, as I said before, Google Analytics or Google Play Store, and they are just not using it right now. We can create some kind of little infrastructure to collect the data from all this data sets that I said, for instance, here also the transaction systems that they already have. You don’t need to export all the data from those transaction system. You can get the most important parts of the data at first to create a simple start. After that, we’ll help them to create a separated data lake using modern technology that doesn’t have big upfront investment. As you get you gain confidence in this approach, you can scale, you can add more infrastructure, you can do a [inaudible], you’ll have the elasticity that you need to scale.


13:35

Here, in the growth part of this diagram, we execute. We start by executing a first experiment for them. Let’s say, for the customer, I’m going to show the case. We started doing it by trying to understand the onboarding process. After we created the first experiment, we were able to talk to the product manager, and help him understand that it was important to also add for the next iteration story or backlog item to handle or to tackle that kind of situation. He agreed, because using the business indicators or some some sort of data, we could show him the value of doing that. As soon as the project continues, and we can start shifting our or mindset from a project perspective to a product perspective, because the product manager or everyone that is involved in the product start to see the value of this kind of [inaudible] thing being developed each iteration.


14:51

So two real use case for customers, I already mentioned some of them before. The first one is their level. A level is a really big company at the financial market. They offer to the market this solution to manage all the benefits our company gives to its employees. For instance, they have cards where the employees can pay for the restaurants, or their grocery shop, and fuel their vehicles, for instance. They have millions and millions of users here in Brazil. They also have a mobile app and different portals. There are all the ecosystems of customers interact with the platform. As I said before, we started simple by adding the first indicator, a first kind of indicator to them. Nowadays, we have this complete set of dashboards of indicators, that the whole team and the whole product areas of the company, follow to understand if we’re going to the right direction with our products with our development.


16:12

As I said before, things like understanding the reviews of the users at the Apple Store or the Play Store. Maybe [unintelligible] not a rocket science, but most of the time, we’re doing and developing mega log items that we’re not dealing with the problems that the users were having actually. We use technologies like Google Cloud for [inaudible] here, Spark for processing lots of data, billions and billions of data. We did a lot of user journey mapping. Here, we have one data scientist working together with the user experience expert. They are all collaborating to create this growth mindset for the product. We also use some techniques of natural language processing to process all the data from the stores, as I said, and a lot of analytics and statistical tools to combine those data, those reviews of this from the stores with the actual transactions wwith the interactions that the customers have read applications.


17:27

Secondly, this is another important customer that we have. They offer satellites TV to the whole country. They have millions and millions of users. They actually have this call center and also their service desk with a chat interface, where other users can contact them, can file some complaints, or requests for solutions. They want to have a better understanding of their users. So in this case, we started by collecting all the calls that the call center had. We created some tools using open source technology to transcribe the audios of the calls to text. We combined those texts with the transcriptions from the chats. Using the same idea that we used for the stores, we could organize the data, this reuse amount of data to segment information, and understand the most important reasons, the most important motives that their customer had when they contact the call centers.


18:46

This was really important because with this understanding, this customer was able to reduce the whole monitoring operation of their call center by 80%. More important than that, we did understanding they are able to feed other areas of the company and other teams with information that were important for them to create better solutions to tackle the problems or to offer the things that our customers were willing to have in their platforms. So here, we use a lot of machine learning techniques, advanced analytics, and some big data infrastructure that are available easily on the cloud nowadays.


19:41

So, this was it. Just a recap of what I talked about in the session. First of all, we need to find better ways to measure the progress of a product development. We can start by creating this process where we move from a project mindset to a product mindset. We start to run into think about our sprints in our product backlog as a series of experimentations instead of a series of predefined features, okay? We can start simple by adding this kind of approach to resist an existing operation that we have right now. That may be using some sort of top down back logs or something like that. You can start simple and showing the value. By showing the data, you can gain the confidence of the product managers, the business owners, and so on. By doing that, we may impact in a very positive way the overall digital strategy of our companies.


20:56

I hope you enjoyed the session. It was a pleasure to meet to talk here at CIO VISIONS. Have a great day.


Get full Q/N Access

Sign up to Q/N with a few details to watch this presentation.

  • Hidden
  • Hidden